KnowVertical Blog

AdSense - Understanding How It Works...

Many bigger publishers or broad network entrepreneurs come into the planning around their network with one big mistaken assumption.  This assumption has existed in the market now for many years, and when people are presented with the truth they are usually shocked.  The problem with this ongoing myth is it is getting more and more off base from reality and has a real risk of having your business models for your networks far off of reality.

Over the past few years this misconception has had ripple effects on sites and on network operators and in many cases has had a dramatic impact on their business plans and financial models.  What is the misconception you might say?  Okay, enough of the drama!

The biggest misconception in the market today has to do with Google AdSense.  Most people in big sites or with limited multi-category experience with AdSense have strong beliefs that the effective CPM paid by Google is extremely low and way below the market rate would be for the similar ad spaces on an open market. 

Most large publishers have the belief that small sites are making pennies on the dollar with AdSense and  that even with a CPM less than a dollar that the network operator would be blowing out the marketplace and publisher expectations. These assumptions make sense, why would random text ads on my blog about my kids be really worth anything to a direct response advertiser?  Why would random text ads showing up along side the video or funny joke I sent my buddies really bring more than a few pennies to the hosting website?  But this is not what you are going to build your ad network on, most likely, and Google AdSense was developed to match advertiser’s goals to consumers.

Yes, there are many cases where Google is paying a ridiculously low rate for their member sites text and image ads.  Rates in the .10 to .25 range for CPM's is not out of the norm for sites that have heavy traffic with low specificity of content (aka focus).  There is a general AdSense effect on pricing of the ad space, depending on a few variables. 

The most important variables in how well AdSense performs are how specific or targeted the content is on the site.  This doesn't translate to sites just about Mesothelioma (the top paying keyword for years on Google); it actually includes all major high value categories online, such as Travel, Business, Financial and more.  This effect also holds true for highly targeted content, for example a site about bed and breakfast hotels in upper Maine targeted toward families with kids would potentially get very high payout due to the very specific content on the site. 

The second variable that has a pretty dramatic effect on how well Google AdSense will pay is how big and how many pages the site has (overall traffic).  The more users and the more pages the less effective the overall advertising will perform on the site.  Though Google has over one million small businesses buying on their keyword product, there is still the potential to run out of high performing/paying ads for the site. 

So if you are recruiting a broad category network that has a lot of general keywords, such as viral entertainment sites, and many of the sites are more vanilla in content, the Google AdSense program will probably be on the very low end of performance.  Most of these sites will also be in other remnant programs to help supplement their revenue, thus even diluting the effectiveness of the AdSense ads more.   

Check out the other post on some of the remnant eCPM's that are currently being aid in the market.  Though the marketplace is constantly changing, the base rules have been consistent for years.  The main message is not to get stung with the AdSense surprise.  The best thing to do when working on an overall model for the network you are building is to do some more detailed research into your marketplace.  Don't rely on just your sites performance on AdSense, as you will be surprised when you realize that some of the sites you are interested in recruiting into your network are making $50+ eCPM in AdSense. 

What we suggest you do if you run into a site that has a high AdSense payout (eCPM), have the site create a new ad unit for your brand sales.  Though you might have to negotiate a different placement, it at least gets them into your sales kit and your network directory.  Remember that the key to a successful network launch is to find creative solutions for the market.

Many bigger publishers or broad network entrepreneurs come into the planning around their network with one big mistaken assumption.  This assumption has existed in the market now for many years, and when people are presented with the truth they are usually shocked.  The problem with this ongoing myth is it is getting more and more off base from reality and has a real risk of having your business models for your networks far off of reality.

Over the past few years this misconception has had ripple effects on sites and on network operators and in many cases has had a dramatic impact on their business plans and financial models.  What is the misconception you might say?  Okay, enough of the drama!

The biggest misconception in the market today has to do with Google AdSense.  Most people in big sites or with limited multi-category experience with AdSense have strong beliefs that the effective CPM paid by Google is extremely low and way below the market rate would be for the similar ad spaces on an open market. 

Most large publishers have the belief that small sites are making pennies on the dollar with AdSense and  that even with a CPM less than a dollar that the network operator would be blowing out the marketplace and publisher expectations. These assumptions make sense, why would random text ads on my blog about my kids be really worth anything to a direct response advertiser?  Why would random text ads showing up along side the video or funny joke I sent my buddies really bring more than a few pennies to the hosting website?  But this is not what you are going to build your ad network on, most likely, and Google AdSense was developed to match advertiser’s goals to consumers.

Yes, there are many cases where Google is paying a ridiculously low rate for their member sites text and image ads.  Rates in the .10 to .25 range for CPM's is not out of the norm for sites that have heavy traffic with low specificity of content (aka focus).  There is a general AdSense effect on pricing of the ad space, depending on a few variables. 

The most important variables in how well AdSense performs are how specific or targeted the content is on the site.  This doesn't translate to sites just about Mesothelioma (the top paying keyword for years on Google); it actually includes all major high value categories online, such as Travel, Business, Financial and more.  This effect also holds true for highly targeted content, for example a site about bed and breakfast hotels in upper Maine targeted toward families with kids would potentially get very high payout due to the very specific content on the site. 

The second variable that has a pretty dramatic effect on how well Google AdSense will pay is how big and how many pages the site has (overall traffic).  The more users and the more pages the less effective the overall advertising will perform on the site.  Though Google has over one million small businesses buying on their keyword product, there is still the potential to run out of high performing/paying ads for the site. 

So if you are recruiting a broad category network that has a lot of general keywords, such as viral entertainment sites, and many of the sites are more vanilla in content, the Google AdSense program will probably be on the very low end of performance.  Most of these sites will also be in other remnant programs to help supplement their revenue, thus even diluting the effectiveness of the AdSense ads more.   

Check out the other post on some of the remnant eCPM's that are currently being aid in the market.  Though the marketplace is constantly changing, the base rules have been consistent for years.  The main message is not to get stung with the AdSense surprise.  The best thing to do when working on an overall model for the network you are building is to do some more detailed research into your marketplace.  Don't rely on just your sites performance on AdSense, as you will be surprised when you realize that some of the sites you are interested in recruiting into your network are making $50+ eCPM in AdSense. 

What we suggest you do if you run into a site that has a high AdSense payout (eCPM), have the site create a new ad unit for your brand sales.  Though you might have to negotiate a different placement, it at least gets them into your sales kit and your network directory.  Remember that the key to a successful network launch is to find creative solutions for the market.

Many bigger publishers or broad network entrepreneurs come into the planning around their network with one big mistaken assumption.  This assumption has existed in the market now for many years, and when people are presented with the truth they are usually shocked.  The problem with this ongoing myth is it is getting more and more off base from reality and has a real risk of having your business models for your networks far off of reality.

Over the past few years this misconception has had ripple effects on sites and on network operators and in many cases has had a dramatic impact on their business plans and financial models.  What is the misconception you might say?  Okay, enough of the drama!

The biggest misconception in the market today has to do with Google AdSense.  Most people in big sites or with limited multi-category experience with AdSense have strong beliefs that the effective CPM paid by Google is extremely low and way below the market rate would be for the similar ad spaces on an open market. 

Most large publishers have the belief that small sites are making pennies on the dollar with AdSense and  that even with a CPM less than a dollar that the network operator would be blowing out the marketplace and publisher expectations. These assumptions make sense, why would random text ads on my blog about my kids be really worth anything to a direct response advertiser?  Why would random text ads showing up along side the video or funny joke I sent my buddies really bring more than a few pennies to the hosting website?  But this is not what you are going to build your ad network on, most likely, and Google AdSense was developed to match advertiser’s goals to consumers.

Yes, there are many cases where Google is paying a ridiculously low rate for their member sites text and image ads.  Rates in the .10 to .25 range for CPM's is not out of the norm for sites that have heavy traffic with low specificity of content (aka focus).  There is a general AdSense effect on pricing of the ad space, depending on a few variables. 

The most important variables in how well AdSense performs are how specific or targeted the content is on the site.  This doesn't translate to sites just about Mesothelioma (the top paying keyword for years on Google); it actually includes all major high value categories online, such as Travel, Business, Financial and more.  This effect also holds true for highly targeted content, for example a site about bed and breakfast hotels in upper Maine targeted toward families with kids would potentially get very high payout due to the very specific content on the site. 

The second variable that has a pretty dramatic effect on how well Google AdSense will pay is how big and how many pages the site has (overall traffic).  The more users and the more pages the less effective the overall advertising will perform on the site.  Though Google has over one million small businesses buying on their keyword product, there is still the potential to run out of high performing/paying ads for the site. 

So if you are recruiting a broad category network that has a lot of general keywords, such as viral entertainment sites, and many of the sites are more vanilla in content, the Google AdSense program will probably be on the very low end of performance.  Most of these sites will also be in other remnant programs to help supplement their revenue, thus even diluting the effectiveness of the AdSense ads more.   

Check out the other post on some of the remnant eCPM's that are currently being aid in the market.  Though the marketplace is constantly changing, the base rules have been consistent for years.  The main message is not to get stung with the AdSense surprise.  The best thing to do when working on an overall model for the network you are building is to do some more detailed research into your marketplace.  Don't rely on just your sites performance on AdSense, as you will be surprised when you realize that some of the sites you are interested in recruiting into your network are making $50+ eCPM in AdSense. 

What we suggest you do if you run into a site that has a high AdSense payout (eCPM), have the site create a new ad unit for your brand sales.  Though you might have to negotiate a different placement, it at least gets them into your sales kit and your network directory.  Remember that the key to a successful network launch is to find creative solutions for the market.

Third Year is a Charm...

The business model is solid.  The CPM's are high.  The sites are great quality.  What does the revenue look like?  It is hard to say for sure how much topline revenue these vertical ad networks really drive, but a deep dive into some of the more established networks shows some good signs.  So to find what the opportunity might be we needed to use some public data and triangulate around total value.  The goal of this exercise was two fold, first to try and better understand how big these networks actually were; but secondly, how well our growth rate predictions fit the vertical ad network model.

To get a good sense of how vertical ad network revenues grow the best proxy is looking at the actual sales from these vertical ad networks.  We built a model that includes the growth trajectory for some of the top vertical ad networks in the market today by using public data from the top audience measurement firms.

To build this analysis I took the traffic and total active site count data for Glam, FM Publishing, and JumpStart networks, these networks have much more publicly available data on the composition of their network, than others.   The next step was to run reports in Media Metrix, Compete, Quantcast, and a few others on the total impressions served for each network, and contributing site.  I triangulated the 3-4 data sources to come up with average impressions count and total unduplicated UV’s for each network.   Of course if possible I relied more on comScore or NetRatings vs the others.

Now, the problem with these numbers is really driven off the fact that impressions alone don’t really tell you much more than how big the potential network is.  The page views or impressions metrics within all the audience measurement services give an estimate of scale, but don’t really tell you anything about sell-through rates or CPM’s.  There are plenty of rumors going around that the sell-through rates on Glam and FM Publishing are low, that is probably true if you measure sell-through as a percentage of all inventory.  But if you calculate the sell-through against the high quality (above the fold, homepage, premium content) channels/pages, then the sell-through on these networks is much higher.

So, to make this more complicated, we dropped the impression growth rates into a time aligned curve  built off of over 100 networks from the beginning of time (okay, since 2000).  This historical growth rate then drove the sell-through rate across these networks; we were making the assumption at this point that these networks will sell their inventory at the rate of the historical trends across the other networks.  We already have their site and reach growth rates, so all we needed was the total impressions on top of the estimate of sell-through.

Using this model, the actual impressions on the network is a proxy for the available inventory to sell.  In the case of FM Publishing and JumpStart, this actually is an accurate assumption because of the way their networks actually operate.  Glam is a slightly different animal and there is a potential within the Glam estimates that their total revenues in 2008 and 2009 might be slightly over stated.  The only reason we don’t think this is a big issue is because Glam is still growing their network, which we are not taking into account in the network, and Glam also has a very good remnant deal with Google, which would then support the higher sell-through across their networks, albeit at a slightly lower eCPM.

Finally, to align and calculate the actual revenue, we used the average CPM’s that our market intelligence tells us is the appropriate rate for the categories these networks are selling into.  The CPM estimate was built based on taking the category CPM’s and then looking at the composition of these networks by the categories they are representing.  Additionally, we have the actual reported/public CPM’s from each of these networks and can make some assumptions on discounting based on rate-card and volume estimates.

The final outcome of this analysis shows a pretty solid growth rate across these three vertical ad networks.   The estimated revenues for 2007 were: Glam  $21MM; JumpStart $59MM; FM Publishing $24MM.  What is the over under on these numbers?  Who knows, but the data seems to be in the range of acceptable error.

The growth rates across these networks ranged from 100% growth to 700% growth.  This growth is not just a function of the model, but really a function of the size of the network and where the network actually is in the model.  Because Glam continues to recruit sites on a monthly basis, their sell-through rate is actually lower because the composition of their network is expanding as they sell more.  JumpStart is actually counter to this as they have a limited set of sites they are working with, at least until they start to recruit more sites, and they are somewhat bound to a doubling of revenue year over year.

These numbers seem very reasonable and if you compare to publicly available data the numbers are within the range of error. 

We believe these networks, in their 3-4th year of operations are going to enjoy significant growth in total revenue, but also higher CPMs and bigger budgets within their category.

Remember these are estimates and the public data that was used to validate this data is probably subject at best.  Please use this analysis with a grain of salt.

A Real Market Trend…

Women, women, everywhere.  No, this is not Tom Brady’s hotel lobby.  Yes, we have seen at least 10 new networks over the past few months.  Almost every one of these new vertical ad networks focused on women, and any given combination of this demographic segment.   From Warner Brothers MomLogic Mom’s Network, to Nickelodeon, to Martha Stewart, and so on.

Yes, advertisers are very interested in reaching women, especially in the 24-45 age groups.   Big CPG, health and beauty, financial and even auto manufactures are targeting women these days, with big budgets and targeted creative.   This year will also be very big in the election, as any political pundit would tell you, women are the elusive swing voters.

So women are a great target online, and there are plenty of options.  But what about the men?  Is there a really effective way to buy across the long tail of high quality sites and reach men?  There are very few vertical ad network options for male oriented advertisers.  Yes, there have been a few that have a male skew, with business and sports, but nothing that directly targets this demographic group.

US men age 18 to 34 comprise nearly 27% of the total male Internet user population, but they are elusive. They continue to watch fewer hours of TV and spend more time online, playing games, downloading videos, music and podcasts and sometimes shopping.

Data from eMarketer reports that the male demo, younger than 45, is the most important target audience today (tied with women) for online marketers.

A quick analysis of the top advertiser categories, based on data from Nielsen and comScore, show the men’s category (i.e. men’s health, entertainment, fashion & lifestyle, living) as #8 in the total available quality inventory for brand ad spending.

 

This demographic is very important and there is plenty of online brand dollars allocated to these consumers.  As the women’s online land grab gets to a heated pace, the men’s sites are just starting to see some real interest, and if the women’s category is an indicator, this energy will become exponential as many players enter the market.

This target is ripe for the picking, and those that execute and have first mover advantage will win!

A Case For Below the Fold…

Most ad networks really want their member sites to ensure the ad placement is above the fold (so no scrolling is necessary).  This is an old concept and has been carried over and over throughout the years.  Most big budget campaigns are looking for above the fold inventory and continue to pay a premium for this placement.

This logic makes sense if you think about CPM based campaigns, because you are not guaranteed an impression if you are below the fold.  If a visitor doesn’t scroll down on the page, you have paid for a wasted impression to the user.

In the world of direct response advertising this is even more important, because you not only want the user to see and engage with the ad, but also click the ad.  If most of your ads are near the bottom of the page, you could have a hard time getting clicks.   For this reason, many marketers in the direct response area have focused on more performance based marketing, via CPC or CPA type payouts.   This protects them in a blind network, such as AdSense; to ensure they are really only paying when there is an event.  The usage of the performance based model can protect the advertiser.

But this line of thinking is moving into the vertical ad network space for brand advertisers.  Many RFP’s  for big budget campaigns are pushing for premium placement, aka above the fold.   The thinking is similar to that of direct response advertisers, you want to pay for an impression that someone sees and engages with.  The problem with this thinking is it actually is not the best way to approach a brand building campaign.

Unlike direct response ads, the brand building campaign is trying to reach engaged visitors to high quality vertical content sites.   The brand message wants to reach their target audience when their defenses are down and their trust is up.  The brand advertiser wants to compliment the overall message with the user experience on a give website or network of sites.   The direct response advertiser wants to reach the user when they have not found everything they were looking for and need more, thus clicking on an ad.

So, within the brand world, the goal is to have an integrated voice and message that is presented to your target customer at the right time and in the right place.  The users that are engaged with a high quality site, within your target niche, are the most attractive target for you and are the most open to interacting with your message in their experience.

Given these goals, do you think that above the fold is really that valuable?  Yes, it is nice to ensure every visitor to a site you are buying media on is going to see your message.  But think about the engagement part of your goal.  Think about the user having their defenses down and being open to your message.  Common sense will tell you that when you visit a site that you are engaged with, you don’t just look above the fold and then leave.  For users that are actually interested in the content of the site, and thus potentially more interested in what you have to say, are going to take the time to scroll down and interact with what the site/blog has to offer.

Yes, direct response is happy to be first if your face when you hit a site that doesn’t offer you what you are looking for.  They are hoping that you will be interested enough, as you are looking for the next place to visit, to click on their ad.  True, these visitors, that have come to a site and are not happy with the content or it is not what they expected, are much more likely to click on a direct response campaign that looks like it is going to bring them closer to what they were searching form.

In the brand advertising world, it is much more important to ensure you are presenting your message to the right customer or prospect.  You are not trying to offer an alternative to the user that is visiting the high quality vertical site/blog.  You are trying to compliment their visit with more information that they will take with them and potentially act upon.

It is also common sense that as users scroll down a site, and spend more time on a site, they are more willing to take in all the information on the site.  Additionally, when these sites or blogs are already within their web trust family, the user is even more interested in what is being presented, and your ad campaign will offer the user a much more valuable experience.

Common sense should drive what you do as you build out your vertical author community.   Think about your target customer and think about where you want them to be, both physically and mentally, when you present your idea or message to them.   More often than not, you want them where they are most open to engage.

Finally, beyond common sense, there was a study just released by ClickTale, one of the emerging Web 2.0 research firms on the web.  They have release a few really compelling studies around online consumer behavior. Their most recent study, Visibility and Scroll Reach, is available on their blog.  They have also released studies on consumer interaction with AJAX and even additional research on site visitation behavior.

The recent study by ClickTale provides a good understanding of the real usage on websites and blogs.  What ClickTale has uncovered, validating our common sense, is that users on a given website or blog are Boolean.  They are either interested in the content on the site, thus engaged and trusting, or not.  The study estimated that more than 75% of users scrolled below the fold in any given visit.  Additionally, the data showed that regardless of page size, the distribution of scrolling was similar.  This last finding is what proves our theory.

Based on the ClickTale study, and our hypothesis, the users that find a site that is interesting are not deterred by the size of the page or amount of scrolling required to engage with the content.   Some of the conclusions from this study point out this exact finding…

It appears that visitors scroll in a relative way - about the same percentage of page views will reach the middle of a web page regardless of the actual page height in pixels.

Visitors appear to be using the location of the scroll bar but not the size of the tracker when scrolling, since the scroll bar location is a relative indicator and the scroll tracker size is an indicator of page height.

Feel free to read the whole study, and many of their other studies at the ClickTale Blog.  And don’t forget to scroll to the bottom; the best part of the study is below the fold.

The Time Has Come...

A big part of building a vertical network is the syndication of video.  If you have a catalog of high quality endemic video, you can quickly build a vertical community.

What this means is you can take your video catalog and distribute it across the web within your network of sites.  This is not the traditional widget approach where you let anyone pick-up the widget and present it, this focuses on premier video that is only distributed to your partner sites.

You can actually increase your distribution in a controlled manner, while ensuring you can monatize this video content at the highest CPM.

The stats are in and online video watching has officially become an obsession, according to comScore's July online video viewing results with nearly 75% of all U.S. internet users watching online video. Online viewers watched an average of 68 videos during the month, consuming more than three hours of viewing time. The average clip spanned 2.7 minutes in length.

Top U.S. Online Video Properties by Unique Video Viewers July 2007
Property                 Unique Viewers(000)            U.S. Internet Reach
Google Sites                   67,782                                 37.6%
Fox Interactive Media     35,834                                 19.9%
Yahoo! Sites                   35,325                                 19.6%
Time Warner Network     26,571                                 14.8%
Viacom Digital                22,652                                 12.6%
Microsoft Sites               18,847                                 10.5%
Disney Online                 13,907                                   7.7%
ESPN                               7,733                                  4.3%
MLB.com                         7,671                                  4.3%
Photobucket.com             6,684                                  3.7%
Source: comScore Video Metrix